312099 - {inc}Text jumps without a cause, when clicking on the button

Status: RESOLVED FIXED

(Core :: Layout: Block and Inline, defect)

Description

[Martijn Wargers (dead)]

This came from: http://forums.mozillazine.org/viewtopic.php?t=328340&highlight=

when clicking on the button on that page, the whole text shifts downwards, that
should not happen.

Comment 1

[Martijn Wargers (dead)]

Attachment: testcase

Comment 2

[Peter van der Woude [:Peter6]]

It works as it should if the table is removed or gets style{width:XXpx} which
has to be wider than div #left + div #right

Comment 3

[Bernd]

Attachment: reflow log

Comment 4

[Bernd]

During the button down the outline causes an incremental reflow,
where the block returns a wrong maximumWidth at the end.

          block 0322CFB0 d=5235,1530 me=3705 m=0 

So what does this line say?
The block would like to occupy 5235 twips (d = desired width), and it should not
be shrinked below 3705 (me = MEW) and if the block had infinite space it would
occupy 0 twips!!! (m =maximumwidth)
I would say the incr. reflow with a float uses a block reflow code path which
does not update the maximumWidth, while it has been requested to do so from the
parent table cell.
The table cell is used to bogus .mMaximumWidth values, see line 996 in
nsTableCellFrame.cpp

// make sure the preferred width is at least as big as the max element width
if (aDesiredSize.mComputeMEW) {
  aDesiredSize.mMaximumWidth = PR_MAX(aDesiredSize.mMaximumWidth,
aDesiredSize.mMaxElementWidth);
}

and wraps to the MEW to mitigate the damage.

Comment 5

[Bernd]

Attachment: testcase whithout interaction

The testcase just causes an incr. reflow without the outline mechanism.
Martijn, it would be interesting to see when it really regressed.

Comment 6

[Bernd]

sending it over to block layout where IMHO it belongs.

Comment 7

[Martijn Wargers (dead)]

(In reply to comment #5)
> Martijn, it would be interesting to see when it really regressed.
Well, it really regressed between 1.3a (2002-12-12) and 1.3 (2003-03-12)

Comment 8

[Bernd]

I would bet a beer on bug 172896

Comment 9

[Martijn Wargers (dead)]

Seems related to bug 246669.

Comment 10

[Peter van der Woude [:Peter6]]

Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9a1) Gecko/20061208 Minefield/3.0a1 ID:2006120812 [cairo]

Seems fixed after reflow branch landing

Comment 11

[Martijn Wargers (dead)]

Yeah, fixed by the reflow branch landing.
Both testcases indeed showed the bug in the 2006-12-06 build, but not anymore in the 2006-12-08 build.

Comment 12

[Ryan VanderMeulen [:RyanVM]]

Adding in-testsuite? nomination per bz's request in m.d.t.l. Sorry for the bugspam.